Statistics::Distributions, error "Invalid n:"

Win has asked for the wisdom of the Perl Monks concerning the following question:

Dear Monks,

The following bit of Perl:

use strict;
use DBI;
use DBD::ODBC;
use Statistics::Distributions;

### A load of code here
### And then:

 while ( my @columns = $sthZL->fetchrow() ) {
        print "    Row " . $count++ . ":" . join("\t",@columns) . "\n"
+;
        
        my $degrees_of_freedom_in_numerator = 2*($columns[4]-$columns[
+6]+1);
        my $degrees_of_freedom_in_denominator = 2*($columns[6]);
        my $fprob=Statistics::Distributions::fprob($degrees_of_freedom
+_in_numerator, $degrees_of_freedom_in_denominator, 0.025);
        
        my $result = ($columns[3]+(sqrt(($columns[6]/1000)/$columns[4]
+)*($columns[4]/$columns[4]+($columns[5]-$columns[4]+1)*$fprob*$column
+s[4]-$columns[3]))*100000);
        print "\n$result\n";
[download]

Gives me:


<code>
    Row 1:All   All     459     7236    286820  3000    .0103270000000
+00000000000000
Invalid n: 573641.979346
[download]

Can anyone please explain 'Invalid n' error message? _{janitored by ybiC: Retitle from "Invalid n - error message" for better searching}

Comment on Statistics::Distributions, error "Invalid n:" Select or Download Code

Replies are listed 'Best First'.
Re: Statistics::Distributions, error "Invalid n:" by idsfa (Vicar) on Feb 04, 2004 at 16:33 UTC
Did you look at the docs for Statistics::Distributions? Or the code? If you had, you would see that fprob only accepts positive integer values for the degrees of freedom (as it should). `sub fprob { # Upper probability F(x,n1,n2) my ($n, $m, $x) = @_; if (($n<=0) \|\| ((abs($n)-(abs(int($n))))!=0)) { die "Invalid n: $n\n"; # first degree of freedom } if (($m<=0) \|\| ((abs($m)-(abs(int($m))))!=0)) { die "Invalid m: $m\n"; # second degree of freedom } return precision_string(_subfprob($n, $m, $x)); }` [download] If anyone needs me I'll be in the Angry Dome.	[reply] [d/l]
Re: Statistics::Distributions, error "Invalid n:" by Old_Gray_Bear (Bishop) on Feb 04, 2004 at 16:29 UTC
It has been a while since I did semi-serious statistics. But, shouldn't the 'degrees of freedom' be intergers? The DoF in the denominator works out to 2*0.10327.... ---- I Go Back to Sleep, Now. OGB	[reply]